Depth Measures for Multivariate Functional Data: an application to ECG signals

Speaker: Francesca Ieva
Affiliation: Politecnico di Milano, MOX - Dipartimento di Matematica,
When: Thursday 27th October 2011
Time: 15:00:00
Abstract: The statistical analysis of functional data is a ever growing interest research area. In biomedical context examples of complex output to be studied with suitable statistical techniques are more and more frequently appearing. The seminar is focused on the analysis of ECG signal (both arising from real clinical observation and numerical simulations) as multivariate functional data. Two are the main focus of the analysis: the application of new multivariate functional indexes of depth to outlier detection, and classification of these signals in unsupervised classification framework in order to get a semiautomatic diagnostic tool for cardiovascular diseases like Bundle Brunch Block.